Nearly 1 in 5 teens and young adults turn to AI chatbots for mental health advice, study finds

Nearly 1 in 5 American adolescents and young adults are now turning to artificial intelligence chatbots for emotional support — a sharp increase from just a year ago — and most are doing so without telling anyone, according to new research that is raising alarms among mental health professionals.

Use of AI chatbots for mental health advice among young people ages 12 to 21 rose by more than 40% over the past year, with 19.2% now reporting they have sought help from tools such as ChatGPT, Gemini, Character.AI and Meta AI when feeling sad, angry, nervous or stressed, up from 13.1% in a similar survey conducted a year earlier. The findings, from the RAND Corporation, were published in JAMA Pediatrics.

Researchers estimate the figure represents roughly 8.2 million young people nationwide, suggesting AI chatbots are becoming an increasingly common source of emotional and mental health support as the United States continues to face a youth mental health crisis. 

The study found that use was widespread even among young people already engaged in formal care, suggesting chatbots are functioning not only as a substitute for professional help but as a supplement to it. Four in 10 teenagers with a major depressive episode in the past year report receiving no mental health services, according to figures cited in the JAMA Pediatrics paper, underscoring the gap AI tools have moved quickly to fill. 

Yet experts warn the tools carry serious risks. Among those who reported using AI chatbots for mental health advice, nearly two-thirds — 63% — said they had not disclosed that use to anyone, while nearly 43% said they sought such advice at least monthly. 

“Many young people appear to be using AI chatbots for mental health advice privately, without the knowledge of parents, clinicians or other adults,” said Jonathan H. Cantor, a RAND senior policy researcher and study co-author. “That makes it especially important for adults to start conversations about how AI tools are being used and the role they should and should not play.”

A large majority, 92%, of users said the advice they received was somewhat or very helpful — though researchers cautioned that may reflect chatbots’ tendency to flatter users rather than the actual quality of the guidance provided. 

The National Academy of Medicine convened a panel of experts earlier this year that reached similarly cautionary conclusions. Panelists warned that AI chatbots may be uniquely capable of causing harm compared to other digital platforms because they simulate an actual relationship and its associated emotions, rather than merely delivering information to consume. Adolescents, they said, are especially vulnerable to the emotionally counterfeit bonds chatbots can create because they are in a sensitive period of brain development, and the tools’ most prominent behaviors — being highly affirming, offering excessive praise, providing one-sided advice — are particularly addictive to teenagers and obstruct development of the skills they need as adults. 

Panelists were direct that no validated therapeutic alternative yet exists.

“There is absolutely no consensus in the field that AI chatbots can serve in any way as a replacement for therapy,” said C. Vaile Wright, senior director of the Office of Healthcare Innovation at the American Psychological Association. 

The RAND study found use was more common among females than males and more prevalent among young adults ages 18 to 21 than among younger teens ages 12 to 17. Youth who had spoken with a physician about their mental health in the prior six months were also more likely to report turning to chatbots for advice.